Synthesis and X-ray characterisation of a new polycondensed heterocycle obtained by a novel Mn(III)-mediated cascade reaction of 2-cyanophenyl isothiocyanate

Article abstract of DOI:10.1016/S0040-4020(01)00678-0

2-Cyanophenyl isothiocyanate reacted with Mn(III) acetate in acetic acid or acetonitrile to give fair yields of a new polycondensed heterocycle arising from the joining together of two molecules of the starting isothiocyanate with loss of a CS moiety. The yields were close to 90% when the reaction was carried out in the presence of diethyl malonate. This compound was unambiguously identified by X-ray crystallography. Under the same conditions, 2-(methoxycarbonyl)phenyl isothiocyanate gave a quinazolinimine derivative instead which is likely to arise from cyclisation of an intermediate N,N′-diarylthiourea. The mechanism of formation of the former compound probably involves formation of a N,N′-bis(2-cyanophenyl)thiourea, followed by rearrangement and radical tandem ring closure of the corresponding cyclic imine derivative. This hypothesis is also supported by semiempirical calculations.

Full text of DOI:10.1016/S0040-4020(01)00678-0

TETRAHEDRON
Pergamon
Tetrahedron 57 (2001) 7221±7233
Synthesis and X-ray characterisation of a new polycondensed
heterocycle obtained by a novel MnꢀIII)-mediated cascade
reaction of 2-cyanophenyl isothiocyanate
²
Gianluca Calestani, Laura Capella, Rino Leardini, Matteo Minozzi, Daniele Nanni,p
Romina Papa and Giuseppe Zanardi
Á
Dipartimento di Chimica Organica `A. Mangini', Universita di Bologna, Viale Risorgimento 4, I-40136 Bologna, Italy
Received 11 April 2001;revised 31 May 2001;accepted 20 June 2001
AbstractÐ2-Cyanophenyl isothiocyanate reacted with Mn(III) acetate in acetic acid or acetonitrile to give fair yields of a new poly-
condensed heterocycle arising from the joining together of two molecules of the starting isothiocyanate with loss of a CS moiety. The yields
were close to 90% when the reaction was carried out in the presence of diethyl malonate. This compound was unambiguously identi®ed by
X-ray crystallography. Under the same conditions, 2-(methoxycarbonyl)phenyl isothiocyanate gave a quinazolinimine derivative instead
which is likely to arise from cyclisation of an intermediate N,N0-diarylthiourea. The mechanism of formation of the former compound
probably involves formation of a N,N0-bis(2-cyanophenyl)thiourea, followed by rearrangement and radical tandem ring closure of the
corresponding cyclic imine derivative. This hypothesis is also supported by semiempirical calculations. q 2001 Elsevier Science Ltd.
All rights reserved.
1. Introduction
report the synthesis and X-ray characterisation of a new
polycondensed heterocycle obtained from 2-cyanophenyl
isothiocyanate through a novel cascade reaction initiated
by manganese(III) acetate.
The discoveries of the last two decades have convincingly
located radical chemistry in the set of most powerful
synthetic tools, and organic free-radicals are now playing
a great role in the synthesis of many sorts of organic
compounds. Our group has been studying a peculiar class
of radical species, i.e. imidoyl radicals, for several years,
which are key intermediates for the formation of various
heterocyclic derivatives.1 Recently, we turned our attention
to a-thio-substituted imidoyl radicals. These species can be
generated by addition of either sulphanyl radicals to iso-
nitriles2 or tin-, silicon-, and carbon-centred radicals to
isothiocyanates.3 In three recent papers4 we have described
the syntheses of benzothieno-fused quinoxalines4a,b and
quinolines4c through cascade reactions of a-thioimidoyl
radicals generated by addition of either 2-cyano- and
2-alkynyl-aryl radicals to aryl isothiocyanates,4a,c or
(2-cyanophenyl)sulphanyl radicals to aryl isonitriles.4b
These results prompted us to further investigate the
reactivity and synthetic potential of isothiocyanates under
various radical or electron-transfer conditions. Here we
2. Results and discussion
When the isothiocyanate 1 (1 equiv.) was allowed to react
with manganese(III) acetate (2 equiv.) in acetic acid or
acetonitrile at 608C, the reaction furnished a previously
unreported aceanthrylene derivative 2 in 45±50% yield
(Scheme 1). The structure of 2 was unambiguously
established by X-ray diffraction (Fig. 1).
As can be noted, the skeleton of this compound formally
involves the joining together of two units of 1 with loss of
one CS molecule. It is worth noting that comparable yields
of 2 were obtained both in acetic acid and in acetonitrile,
whereas the presence of an enolisable compound such as
diethyl malonate resulted in shorter reaction times and
Keywords: isothiocyanates;thioureas;manganese(III) acetate;X-ray
diffraction;rearrangement;radicals.
p
Corresponding author. Tel.: 139-051-2093623;fax: 139-051-2093654;
e-mail: nanni@ms.fci.unibo.it
²
Present address: Dipartimento di Chimica Generale ed Inorganica,
Á
Analitica e Chimica Fisica, Universita di Parma e Centro di Studio per
la Strutturistica Diffrattometrica del CNR, Viale delle Scienze, I-43100
Parma, Italy
Scheme 1.
0040±4020/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(01)00678-0

7222
G. Calestani et al. / Tetrahedron 57 /2001) 7221±7233
intermediate in the reaction of isothiocyanate 1 also, we
tried to synthesise the 2,20-dicyano derivative 7. Analogous
to what happened with 3, any attempt to obtain compound 7
by usual methods resulted in isolation of the iminoquinazo-
line 8 in high yields (Fig. 2).10 On the other hand, no traces
of compound 2 were detected under these conditions. The
1
structure of compound 8 was clearly suggested by H and
13C NMR analyses, which showed, respectively, two
distinct N±H signals and the presence of seven quaternary
and eight tertiary carbons: this data are clearly inconsistent
with a symmetrical structure like 7, but they perfectly ®t the
cyclised structure of 8.
Figure 1. Molecular structure of 2, showing the atom labelling scheme.
When imine 8 was treated at 608C with manganese(III)
Scheme 2.
nearly quantitative yields (,90%).5 At this stage it is not
clear how diethyl malonate affects the reaction outcome;
however, its importance most likely resides in its aptitude
to give rise, by Mn(III) oxidation,6 to carbon radicals that
could somehow initiate or promote the cascade reaction.
acetate in acetic acid or acetonitrile and in the presence or
absence of diethyl malonate (DEM), it systematically
afforded the polycyclic compound 2, although in variable
yields. Table 1 shows that acetic acid is a good solvent as far
as the reaction times are concerned;nevertheless, it gave
much lower yields with respect to acetonitrile (entries a and
c). Better yields were also ensured by the presence of DEM
(entries b and d). The absence of the manganic salt caused
instead the virtual disappearance of product 2, even after
5 days at 608C (entry e).
To throw some light on the mechanism of formation of
compound 2, we allowed the methoxycarbonyl-substituted
isothiocyanate 3 to react with manganese(III) acetate
(2 equiv.) in acetic acid at 608C in the presence or absence
of diethyl malonate. In both cases, the reaction gave the
quinazoline derivative 4 in ca. 30% yield, together with
small amounts of N-acetylanthranilate 5 (Scheme 2).7 As
one can see, the reaction outcome is completely different
with respect to isothiocyanate 1;it is worth noting that the
structure of product 4, still containing the thiocarbonyl
moiety, seems closely related to that of an intermediate,
symmetrically disubstituted thiourea, viz. 6.
In light of these data, imine 8 can therefore be assumed as a
Actually, compound 4 was the only product obtained from
any attempt to independently synthesise thiourea 6 from
methyl anthranilate, starting either from iodine and pyridine
in carbon disulphide8 or thiophosgene and pyridine in
benzene.9 Therefore, it may be reasonably inferred that
thiourea 6 is initially formed by treatment of 3 with
Mn(OAc)3 and it then cyclises to 4 through an ionic and/
or radical mechanism.
To test the possibility that a thiourea could be the same key
Figure 2.

G. Calestani et al. / Tetrahedron 57 /2001) 7221±7233
Table 1. Dependence of the yields of 2 on the reaction conditions
7223
Entry
Mn(III) (equiv.)
DEM (equiv.)
Solvent
t (h)
2 (%)
a
b
c
d
e
2
2
2
2
0
0
1
0
1
0
AcOH
AcOH
MeCN
MeCN
MeCN/H2O
14
22
48
48
25
50
70
83
150
Trace amounts
Scheme 3.
viable intermediate for the conversion of 1 into 2. It is
necessary to explain the mechanism of formation of 2
from 8, which entails a rearrangement of the iminoquinazo-
line with migration of the 2-cyanophenyl ring from the
heterocyclic to the iminic nitrogen atom. Such a transloca-
tion would afford the amine 9 (Fig. 2), whose structure is
much closer to that of the aceanthrylene and could quite
easily account for the formation of 2 through a cascade
cyclisation.
requiring 150 h for converting 70% of the starting material.
The same experiment performed in AcOH/water for 48 h
resulted in hydrolysis of 8 to the corresponding ketone
(27%).12 Also the intervention of Mn(OAc)3, together with
its crystallisation water, as an acid catalyst should be
discarded, since the acid catalysis seems to favour the
unrearranged, hydrolysis products. This is proved, besides
the result obtained in acetic acid, by treatment of 8 with
hydrochloric acid in re¯uxing dioxane for 6 h, which
afforded mainly two compounds (19 and 20, Fig. 3) derived
from hydrolysis of the imine (19) and nitrile (20) moieties of
8;compound 21, arising from hydrolysis of the rearranged
ketimine 9, was only obtained in very small amounts (4%)
(for a detailed discussion about identi®cation of compound
20, see Section 4).
As a convincing evidence for the occurrence of a rearrange-
ment, when thiourea 10 was allowed to react under the usual
conditions it afforded amine 12 as one of the major products
(Scheme 3). The structure of 12 was assigned by way of
spectral analogies with an analogous compound, obtained
from thiourea 30a, whose structure was con®rmed by X-ray
diffraction (Fig. 6).
The most viable mechanism that might be envisaged to
account for the migration is a Dimroth rearrangement.11
This process has been usually encountered in 1-alkyl-2-
iminopyrimidines 16, which, in the presence of nucleophiles
(typically water), give an apparent migration of the alkyl
group from the heterocyclic to the iminic nitrogen atom.
Since this mechanism has also been proposed for the
rearrangement of 17- and 18-like substrates (Fig. 3),11 it
could be reasonable to postulate an analogous pathway for
the conversion of imino- (8) into (arylamino)-quinazolines
(9).
On the other hand, in our opinion, the Dimroth rearrange-
ment is not likely to be the precise or the only mechanism
operating in our reactions. Indeed, the reported rearrange-
ments usually occur under very different conditions, i.e. in
alkaline, aqueous solutions. Moreover, although imine 8
yielded amine 9 when kept at 608C in acetonitrile/water in
the absence of Mn(III), this reaction was very slow,
Figure 3.

7224
G. Calestani et al. / Tetrahedron 57 /2001) 7221±7233
Taking into account the overall above results, we can put
forth a hypothesis, supported by literature precedents, that
allows for the possibility that 8-like thioxoquinazolines
might be oxidised by Mn(III) to give thioamidyl radicals.
Indeed, thiourea and N-substituted thioureas have been
known for many years as corrosion inhibitors and redox
components for vinyl polymerisation in the presence of a
variety of oxidants. Several studies concerning the one-
electron oxidation of thiourea derivatives with oxidising
radicals or metal ions such as Ce(IV) or Mn(III) have
been recently reported.13 It is generally admitted that these
derivatives are easily oxidised to give bidentate, resonance-
stabilised radicals with an intramolecular three-electron
bond between sulphur and nitrogen atoms. We can therefore
surmise that imine 8 could be oxidised to radical 22;the
formation of the isothiocyanate moiety would hence be the
result of a homolytic b-fragmentation with release of a
reasonably stable anilino±amidino radical, viz. intermediate
23 (Scheme 4). Rotation around the Ar±CvN bond and
re-cyclisation onto the NCS group would afford the rear-
ranged product 9. This would therefore be the radical
version of the Dimroth rearrangement.
thiocarbonyl groups (Scheme 4, route b) is probably the key
process that shifts the equilibria towards the aceanthrylene
derivative. In the absence of such a route only mixtures of
rearranged and unrearranged compounds and their
hydrolysis products can be obtained. The fact that 2 is
quantitatively obtained with two equiv. of manganic salt
also starting from 1 (rather than 8) suggests that either the
Mn(III)-catalysed cyclisation of thiourea to the correspond-
ing imine is not a radical process or the result of that
cyclisation is a radical species that reacts further to give 2
before being quenched by some atom-transfer reaction. The
latter hypothesis could be supported by the important side-
products always obtained when we treated 8 (instead of 1)
with Mn(OAc)3 (see Section 4). At this stage we cannot say
anything else about the mechanism;the issue of when and
how DEM intervenes in the reaction remains an unsettled
question as well.
To our knowledge, additions of iminyl radicals to the
sulphur atom of thiocarbonyls have never been observed.
Nevertheless, very recently, we observed some homolytic
intramolecular substitutions of iminyl radicals onto the
sulphur atom of sulphide moieties resulting in the formation
of benzoisothiazole rings.14 This can de®nitely be
mentioned as a good support for our hypothesis. Further-
more, it is worth pointing out that only a radical mechanism
can conveniently explain the formation of a nitrogen±
sulphur bond like that of 2.
Besides explaining the rearranged product 9, this
mechanism can go further, giving on its own a full rationale
of the entire process, including the formation of the poly-
cyclic compound 2. Indeed, radical 24, through its meso-
meric form 2400 can cyclise onto the cyano group to afford
iminyl radical 25. Ring closure of 25 onto the thiocarbonyl
yields radical 26, which furnishes the ultimate product 2
through (formal) loss of a hydrogen atom. The ®nal step
most likely involves oxidation of 26 by Mn(III) to the corre-
sponding cation with subsequent loss of a proton;this can
explain the need for two equiv. of Mn(OAc)3, the former for
generation of radical 22 and the latter for the ®nal oxidation.
It is therefore unnecessary to admit the intermediate forma-
tion of compound 9 by hydrogen abstraction from whatever
species (Scheme 4, route a), since this would require two
additional equiv. of Mn(III) to restart the cyclisation
process. The tandem ring closure involving the cyano and
Additional support to the presence of radical mechanisms is
also given by the side-products 14 and 15 (Scheme 3). These
compounds can be easily rationalised by homolytic
aromatic substitution of a thioamidyl radical on either aryl
rings of the starting compound. More precisely, compound
14 can arise from cyclisation of thioamidyl 27 on the cyano-
substituted aryl ring of thiourea 10, whereas quinazo-
linothiophene 15 is the result of ring closure of thioamidyl
28 on the phenyl ring of the intermediate iminoquinazoline
17 (Scheme 5). Alternatively, 15 might also arise from prior
cyclisation of thioamidyl 27 on the unsubstituted ring,
Scheme 4. Radical mechanism of formation of compound 2.

G. Calestani et al. / Tetrahedron 57 /2001) 7221±7233
7225
Scheme 5. Radical side-reactions of thioureido radicals.
followed by ring closure to the eventual iminoquinazoline.
An analogous compound (29) was also obtained in ,30%
yield in the reaction of imine 8 with Mn(III) and DEM in
acetic acid (Table 1, entry b): its structure was con®rmed by
X-ray diffraction.
(20%) obtained with the SMe moiety, clearly stands for a
radical mechanism.
With the aim of ®nding additional support to our hypothesis
of a radical translocation (Scheme 4), we performed
theoretical semiempirical calculations on the intermediates
and transition states involved in the rearrangement of
radical 28. The results are summarised in Fig. 4.
Homolytic aromatic substitution by sulphur-centred radicals
is a reversible process15 and substitution products are
formed only in the presence of either a good radical leaving
group, e.g. halogen atoms16 or arylthio moieties,17 or in a
highly oxidising medium that is able to oxidise the inter-
mediate cyclohexadienyl radical.18 In our case, it is
plausible that Mn(III) could be the oxidising agent for the
cyclohexadienyls, thus shifting the equilibria towards the
cyclisation products.
Since semiempirical methods are not suitable to estimate
absolute activation barriers, we should focus our attention
to the relative quantities that were furnished by calculations.
Those data clearly indicate that once radical 28 has frag-
mented, the resulting intermediate 31 (or 32) can evolve into
the cyclised structure 33 by surmounting an activation
barrier (Eaˆ9.9 kcal/mol) that is 2.5 times lower than that
of the reverse reaction (Earˆ24.4 kcal/mol). Therefore, in
the absence of signi®cant, fast side-reactions, it is con-
ceivable to state that 28 can be totally converted into the
thermodynamically preferred isomeric radical 33. In the
presence of a fast route that can transform the latter
rearranged radical into another intermediate, like in the
case of the formation of compound 2 (Scheme 4), the
reaction is hence supposed to yield products not derived
from 28 but instead from 33. The results of calculations
are therefore perfectly consistent with the proposed
mechanism.
It is also worth noting that the reactions of the N0-substituted
N-(2-cyanophenyl)thioureas 30a,b with Mn(III) showed
that the presence on the N0-ring of ortho-substituents such
as chloro or methylsulphanyl again led to the formation of
compound 15 (Scheme 6): this can be obviously rationalised
by homolytic ipso-substitution of the thioamidyl 28 (or 27)
onto the C±Cl or C±SMe aromatic carbon. This result,
together with the signi®cant increase of the yield of 15
3. Conclusions
Treatment of 2-cyanophenyl isothiocyanate
1 with
Scheme 6.

7226
G. Calestani et al. / Tetrahedron 57 /2001) 7221±7233
Figure 4. Reaction pathway (with calculated intermediates and transition states) for the rearrangement of radical 28.
Mn(OAc)3 in acetic acid or acetonitrile results in formation
of the novel polycondensed nitrogen-heterocycle 2 in fair
yields. The presence of diethyl malonate brings about a
signi®cant increase of the yield (,90%). The formation of
2 probably involves formation of a N,N0-bis(2-cyano-
phenyl)thiourea, cyclisation to an iminoquinazoline, and
an (apparent) aryl translocation that resembles the Dimroth
rearrangement but instead probably entails a radical
mechanism initiated by Mn(III)-oxidation of the thioamido
moiety. Eventually, a radical cascade reaction brings it to
the ®nal heterocycle.
with CDCl3 and the solvent signal with DMSO and MeCN.
Mass spectra (MS) and high-resolution mass spectra
(HRMS) were performed with a VG 7070E spectrometer
by electron impact with a beam energy of 70 eV. IR spectra
were recorded in chloroform on a Perkin±Elmer 257 instru-
ment or in nujol on a Perkin±Elmer Spectrum RX I FT-IR
spectrophotometer. Column chromatography was carried
Ê
out on silica gel (ICN Silica, 63±200, 60 A) using light
petroleum (40±708C) and a light petroleum/diethyl ether
(or dichloromethane) gradient (from 0 up to 100% more
polar component) as eluant. Previously reported reaction
products were identi®ed by spectral comparison and/or
mixed mp determination with authentic specimens.
The high ef®ciency of this process, which involves so many
steps;the totally new structure of the aceanthrylene deriva-
tive 2 with an interesting N±S linkage;the novelty of the
Dimroth-like radical rearrangement;the unprecedented
radical addition of an iminyl to a thioureido group;in the
end, the possibility that mild radical generation from thio-
amido moieties by Mn(OAc)3 discloses novel synthetic
routes;in our opinion, make this reaction remarkable.
Studies are underway to exploit the potentialities in organic
synthesis of the Mn(III)-mediated oxidation of thioureido
compounds and their analogues.
4.2. Starting materials
Thiophosgene, anthranilonitrile, methyl anthranilate,
diethyl malonate (Aldrich), and manganese(III) acetate
dihydrate (Acros) were commercially available.
4.3. Synthesis of the aryl isothiocyanates 1 and 3
4.3.1. General procedure.19 A dichloromethane (25 mL)
solution of anthranilonitrile or methyl anthranilate
(0.1 mol) was added dropwise in 10 min at rt to a dichloro-
methane (15 mL)/water (35 mL) solution of thiophosgene
(14.4 g, 0.125 mol). The mixture was stirred at rt for 3 h.
Additional dichloromethane was added (50 mL) and the
organic layer was separated and dried (sodium sulphate).
The solvent was evaporated and the residue crystallised
(1) or distilled (3) to give the title isothiocyanate.
4. Experimental
4.1. General procedures
Melting points were determined on an Electrothermal
1
capillary apparatus and are uncorrected. H and 13C NMR
spectra were recorded in chloroform-d1 (where not stated),
deuterated dimethylsulphoxide (DMSO-d6), or acetonitrile
(MeCN-d6) on Varian Gemini 200 (200 MHz) or 300
(300 MHz) instruments;the reference was tetramethylsilane
4.3.2. 2-Isothiocyanatobenzenecarbonitrile ꢀ1).19 Yieldˆ
82%;mp ˆ62±638C (pale-yellow needles, from light
petroleum);200 MHz 1H NMR d 7.29±7.41 (2H, m),

G. Calestani et al. / Tetrahedron 57 /2001) 7221±7233
7227
7.55±7.62 (2H, m);50 MHz 13C NMR d 109.87 (q), 115.89
(q), 127.45, 127.73, 133.81, 134.56, 141.06 (q, very broad,
NCS) (one quat. carbon not apparent); nmax 2220 (CN),
2020 (vs, NCS) cm21;MS m/e (rel inten) 160 (M1, 100),
133 (6), 116 (8), 102 (40), 76 (17), 75 (30).
column chromatography of the crude (light petroleum/
dichloromethane 80:20 v/v) yielded 2 (47%), anthranilo-
nitrile (trace amounts), and N-/2-cyanophenyl)acetamide
(31%).
4.4.4. Reaction of 3 with DEM in acetic acid at 608C.
After 24 h (12112 h), column chromatography of the
crude (light petroleum/diethyl ether 85:15 v/v) afforded
methyl 2-[4-oxo-2-thioxo-1,4-dihydro-3/2H)-quinazolinyl]-
benzenecarboxylate (4) (31%), mpˆ238±2408C (from ethyl
4.3.3. Methyl 2-isothiocyanatobenzenecarboxylate ꢀ3).
Yieldˆ80%;bp (6 mbar) ˆ143±1458C [lit.20 bp
(0.35 mmHg)ˆ100±1028C].
1
acetate) [200 MHz H NMR d 3.75 (3H, s), 7.15 (1H, d,
4.4. Reactions of the aryl isothiocyanates with
manganeseꢀIII) acetate
Jˆ7.8 Hz), 7.27±7.38 (2H, m), 7.55±7.70 (3H, m), 8.17
(1H, dd, J1ˆ8.2 Hz, J2ˆ1.0 Hz), 8.24 (1H, dd, J1ˆ7.8 Hz,
J2ˆ1.6 Hz), 10.60 (1H, bs);50 MHz 13C NMR d 52.84,
115.45, 116.87 (q), 125.49, 127.72 (q), 129.24, 129.82,
130.86, 132.48, 134.37, 136.15, 139.38 (q), 139.72 (q),
160.81 (q), 165.09 (q), 176.88 (q); nmax 3400, 3000, 1720,
1700, 1620 cm21;MS m/e (rel inten) 312 (M1, 15), 281 (3),
253 (100), 162 (7), 90 (7). Anal. calcd for C16H12N2O3S: C,
61.52;H, 3.87;N, 8.96. Found: C, 61.80;H, 3.88;N, 9.00.],
methyl 2-/acetylamino)benzenecarboxylate (5, 6%),
mpˆ97±998C (lit.23 mpˆ1018C) [200 MHz 1H NMR d
2.09 (3H, s), 3.77 (3H, s), 6.92 (1H, td, Jtˆ7.2 Hz,
Jdˆ0.8 Hz), 7.39 (1H, td, Jtˆ7.7 Hz, Jdˆ1.7 Hz), 7.88
(1H, dd, J1ˆ7.2 Hz, J2ˆ1.7 Hz), 8.56 (1H, dd, J1ˆ7.7 Hz,
J2ˆ0.8 Hz), 10.91 (1H, bs);50 MHz 13C NMR d 25.95,
52.77, 115.20 (q), 120.75, 122.87, 131.23, 135.11, 142.02
(q), 169.21 (q), 169.53 (q); nmax 3320, 2960, 1710, 1690,
1610, 1590 cm21;MS m/e (rel inten) 193 (M1, 32), 151
(64), 119 (100), 92 (8), 43 (29)], and a solid, mpˆ130±
1328C, whose structure was not fully established but,
since only three aromatic protons were observed, it is prob-
ably the result of some cyclisation of the isothiocyanato
4.4.1. General procedure. A mixture of aryl isothiocyanate
1 or 3 (5 mmol), diethyl malonate (DEM, 5 mmol), and
manganese(III) acetate dihydrate (5 mmol) in acetic acid
or acetonitrile (25 mL) was stirred at the appropriate
temperature (60 or 808C) until the starting brown colour
of manganese(III) turned yellowish. Another equiv.
(5 mmol) of Mn(OAc)3 was then added and the mixture
was stirred until reduction of additional Mn(III) was
complete;at this point all the starting material was
consumed as well. The mixture was poured into water,
neutralised with 10% aqueous sodium carbonate, and then
extracted with dichloromethane. The organic phase was
dried (sodium sulphate), the solvent was evaporated, and
the residue chromatographed on a silica gel column. The
following reactions were performed according to this
general procedure. All of the column chromatographies
separated an initial fraction containing elemental sulphur.
It is worth noting that the starting material was entirely
recovered by treatment of 1 in acetic acid or acetonitrile
at 608C in the absence of Mn(III) and/or diethyl malonate.
1
moiety on the benzenic ring [200 MHz H NMR d 3.98
(3H, s), 7.20 (1H, dd, J1ˆJ2ˆ7.9 Hz), 7.57 (1H, bd,
Jˆ7.9 Hz), 7.88 (1H, dd, J1ˆ7.9 Hz, J2ˆ1.0 Hz), 10.15
(1H, bs);50 MHz 13C NMR d 53.10, 113.16 (q), 122.73,
126.78 (q), 127.75, 127.81 (q), 128.04, 166.92 (q), 170.15
(q); nmax 3370, 3000, 1700 (b), 1600, 1450, 1290 cm21;MS
m/e (rel inten) 209 (M1, 51), 177 (100), 121 (32). HRMS
calcd for C9H7NO3S 209.0147, found 209.0150]: we suggest
the structure of (previously unreported) methyl 2-oxo-2,3-
dihydro-1,3-benzothiazole-4-carboxylate.
4.4.2. Reactions of 1 in acetic acid. After 38 h (18 h for the
®rst equivalent of Mn(III) and 20 h for the second one),
column chromatography of the crude (light petroleum/
dichloromethane 80:20 v/v) afforded 6-thia-5,7,11c,12-
tetraazabenzo[e]aceanthrylene (2) (87%), pale-yellow
solid, mpˆ224±2278C (from ethyl acetate);200 MHz 1H
NMR d 7.27±7.42 (3H, m), 7.55±7.65 (3H, m), 8.13 (1H,
bd, Jˆ7.5 Hz), 8.36 (1H, dd, J1ˆ7.8 Hz, J2ˆ1.4 Hz);
50 MHz 13C NMR (DMSO-d6) d 124.24, 125.20, 126.10,
126.25, 127.14, 127.58, 134.47, 134.76;21 nmax 3000, 1630,
1610, 1590, 1470 cm21;MS m/e (rel inten) 276 (M1, 100),
138 (11), 102 (15), 41 (23). Anal. calcd for C15H8N4S: C,
65.20;H, 2.92;N, 20.28. Found: C, 65.38;H, 2.91;N, 20.24.
See below for the X-ray characterisation of 2.
The same compounds and yields were obtained by reacting
3 with Mn(III) in acetic acid in the absence of diethyl malo-
nate. On the contrary, only the unreacted starting material
was recovered after 48 h in the reaction of 3 with diethyl
malonate in acetic acid in the absence of manganic acetate.
In the absence of DEM the reaction was complete after 64 h,
the yield of 2 dropped to ,50%, and some N-/2-cyano-
phenyl)acetamide (10%) was separated, mpˆ132±1348C
(lit.22 mpˆ1338C). At 808C the reaction was complete
after 10 h (416 h), but chromatography of the crude (light
petroleum/dichloromethane 80:20 v/v) gave 2 in only 20%
yield together with N-/2-cyanophenyl)acetamide (6%).
4.5. Synthesis of N,N0-diarylthioureas ꢀor the
corresponding imines)
4.5.1. General procedure. A toluene (25 mL) solution of
arylamine (20 mmol) was added dropwise at 358C to a
toluene (25 mL) solution of isothiocyanate 1 (20 mmol)
and cat. amount of p-toluenesulphonic acid. The resulting
solution was stirred at 358C for 24 h;then cooled and
®ltered to give the thiourea derivatives as highly insoluble
solids. With 2-aminobenzonitrile the corresponding imino-
quinazoline 8 was obtained instead of the expected thiourea
7.
4.4.3. Reactions of 1 in acetonitrile at 608C. After 38 h
(18120 h), column chromatography of the crude (light
petroleum/dichloromethane 80:20 v/v) afforded 2 (87%)
and anthranilonitrile (trace amounts). In the absence of
DEM the reaction was complete after 64 h (28136 h) and

7228
G. Calestani et al. / Tetrahedron 57 /2001) 7221±7233
4.5.2. N-ꢀ2-Cyanophenyl)-N0-phenylthiourea ꢀ10). Yieldˆ
86%, mpˆ242±2438C (lit. mpˆ165±1688C10a and
.3008C22); nmax (CHCl3) 3400, 3360, 2215, 1585 cm21;
MS m/e (rel inten) 253 (M1, 100), 252 (45), 220 (30), 195
(30), 93 (15), 77 (35).
4.5.6. Synthesis of 2-[ꢀ2-thioxo-1,2-dihydro-4-quinazo-
linyl)amino]benzonitrile ꢀ9). A solution of iminoquinazo-
line 8 (2.78 g, 10 mmol) in aqueous acetonitrile (0.4 mL of
water in 50 mL of MeCN) was kept at 608C for 150 h. The
reaction mixture was cooled, the solvent was evaporated
and the residue chromatographed to give the title
compound, yieldˆ70%, mpˆ324±3268C (dec.);300 MHz
1H NMR (DMSO-d6) d 7.15±8.45 (8H, m), 10.62 (1H, bs,
NH), 11.35 (1H, bs, NH), 12.54 (1H, bs, NH), 12.91 (1H, bs,
NH); nmax (nujol) 3260, 2215, 1625, 1610 cm21;MS m/e
(rel inten) 278 (M1, 34), 277 (100), 139 (9), 102 (8), 76 (4).
Anal. calcd for C15H10N4S: C, 64.72;H, 3.62;N, 20.12.
Found: C, 65.05;H, 3.60;N, 20.26. The 1H NMR spectrum
shows that compound 9 exists as a nearly 1:1 mixture of
isomers arising from tautomerism of the ±NH±CvN±
moiety. The reported spectrum is not a mixture of 8 and
9, since the higher-®eld NH signal of 8 clearly disappeared.
The formation of a new species was clearly evidenced as
well by the signi®cantly different absorptions showed by the
IR spectrum. The structure of 9 was also con®rmed by its
hydrolysis with dioxane/HCl, which yielded 2-thioxo-2,3-
dihydro-4/1H)-quinazolinone (21) in ,70% yield (identi-
®ed by comparison with an authentic, commercial sample
[Aldrich]).
4.5.3. N-ꢀ2-Cyanophenyl)-N0-ꢀ2-chlorophenyl)thiourea
ꢀ30a). The general procedure was slightly modi®ed by
using dichloromethane (25 mL) to dissolve 2-chloroaniline
and a dichloromethane/light petroleum 1:1 mixture (25 mL)
to dissolve 1. Yieldˆ85%, mpˆ240±2438C;300 MHz 1H
NMR (DMSO-d6) d 7.37 (1H, t, Jˆ7.7 Hz), 7.43 (1H, d,
Jˆ8.0 Hz), 7.49±7.64 (3H, m), 7.65±7.79 (2H, m), 8.23
(1H, d, Jˆ8.0 Hz), 12.30 (1H, bs, NH), one NH signal not
apparent;both NH signals seem to be apparent, as a single
peak, in acetonitrile solvent [300 MHz 1H NMR (MeCN-d3)
d 7.33±7.52 (3H, m), 7.55±7.76 (4H, m), 7.79 (1H, dd,
J1ˆ7.7 Hz, J2ˆ1.6 Hz), 8.55 (2H, bs, NH); nmax (nujol)
3331, 3216, 2238, 1638, 1615, 1544, 1466, 1415 cm21
;
MS m/e (rel inten) 289 (M112, 18), 287 (M1, 50), 252
(100), 219 (7), 161 (8), 111 (10), 90 (11). Anal. calcd for
C14H10ClN3S: C, 58.43;H, 3.50;N, 14.60. Found: C, 58.63;
H, 3.49;N, 14.65].
4.5.4. N-ꢀ2-Cyanophenyl)-N0-[2-ꢀmethylsulphanyl)phenyl]-
thiourea ꢀ30b). Yieldˆ85%, mpˆ224±2268C; nmax (nujol)
3254, 3186, 2224, 1575, 1534, 1498 cm21;MS m/e (rel
inten) 299 (M1, 77), 266 (24), 252 (100), 220 (10), 161
(10), 106 (30), 102 (19). Anal. calcd for C15H13N3S2: C,
60.17;H, 4.37;N, 14.03. Found: C, 60.40;H, 4.39;N,
13.99. As soon as compound 30b is dissolved in DMSO,
it appears to cyclise to the corresponding iminoquinazoline;
the following NMR spectrum is consistent with the latter
structure (6:4 mixture of imine geometric isomers):
4.6. Reactions of iminoquinazoline 8
4.6.1. Reaction of 8 with MnꢀOAc)3 in acetic acid ꢀTable
1, entry a). Following the general procedure in Section
4.4.1, after 14 h column chromatography separated 2
(25%), mpˆ224±2268C (spectral data identical to those
reported in Section 4.4.2) and 13H-quinazolino[3,4-
a]quinazolin-13-one (5%), mpˆ112±1158C (dec.) [nmax
(CHCl3) 1693 cm21;MS m/e (rel inten) 247 (M1, 100),
219 (85), 129 (18), 119 (25), 102 (35), 76 (23);HRMS
calcd for C15H9N3O 247.0746, found 247.0752];the latter
compound could not be puri®ed, but its 300 MHz 1H NMR
spectrum (DMSO-d6) clearly showed the signal due to the
H-2 proton (d 8.53, s);the structure was also deduced by its
analogies with compound 29 (see below), which it can
derive from by desulphuration.25 The reaction also afforded
major amounts of an unidenti®able product with (apparent)
m/e 259, which proved to be completely insoluble in all of
the most common solvents.
1
300 MHz H NMR (DMSO-d6) d 6.60 (0.6H, bs, imine
NH), 7.29±7.65 (6H, m), 7.71 (1H, t, Jˆ7.5 Hz), 8.19
(1H, d, Jˆ7.7 Hz), 9.28 (0.4H, bs, imine NH), 12.49 (1H,
bs, thioamide NH);24 attempts to record spectra in less polar
solvents were unsuccessful, due to the extreme insolubility
of 30b. An IR spectrum recorded after the NMR experiment
showed totally different NH and CN stretching regions, thus
con®rming complete conversion of 30b to the correspond-
ing iminoquinazoline: nmax (nujol) 3213, 1636, 1614, 1543,
1414, 1304, 1259, 1229, 1183 cm21. This compound melts
at 224±2268C, so that it is presumable that the mp reported
above for 30b is actually that of the cyclisation product.
4.6.2. Reaction of 8 with MnꢀOAc)3 and DEM in acetic
acid ꢀTable 1, entry b). Following the general procedure in
Section 4.4.1, after 22 h column chromatography afforded 2
(50%) and 11H-4-thia-5,10,11c-triazacyclopenta[def]chry-
sen-11-one (29, 30%), mpˆ295±3008C [300 MHz 1H NMR
(DMSO-d6) d 7.86 (1H, t, Jˆ7.5 Hz), 7.95±8.03 (2H, m),
8.12 (1H, d, Jˆ7.7 Hz), 8.17 (1H, d, Jˆ8.0 Hz), 8.49 (1H, d,
Jˆ7.7 Hz), 8.75 (1H, d, Jˆ7.7 Hz); nmax (nujol) 1657, 1595,
1511 cm21;MS m/e (rel inten) 277 (M1, 58), 249 (100), 217
(7). Anal. calcd for C15H7N3OS: C, 64.97;H, 2.54;N, 15.15.
Found: C, 65.41;H, 2.54;N, 15.19.]. The structure of 29
was con®rmed by X-ray diffraction (Fig. 5).
4.5.5. 2-[4-Imino-2-thioxo-1,4-dihydro-3ꢀ2H)-quinazo-
linyl]benzonitrile ꢀ8). This product was always obtained
by any attempt to synthesise thiourea 7 from 1 and 2-amino-
benzonitrile.
Yieldˆ80%,
mpˆ172±1758C
(dec.);
1
300 MHz H NMR (DMSO-d6) d 7.40 (1H, t, Jˆ7.7 Hz),
7.46 (1H, d, Jˆ8.0 Hz), 7.66±7.78 (3H, m), 7.93 (1H, ddd,
J1ˆJ2ˆ8.0 Hz, J3ˆ1.4 Hz), 8.08 (1H, dd, J1ˆ7.7 Hz,
J2ˆ1.1 Hz), 8.27 (1H, d, Jˆ7.7 Hz), 9.35 (1H, bs, NH),
12.58 (1H, bs, NH);75 MHz 13C NMR (DMSO-d6) d
112.83 (q), 115.14 (q), 115.97, 116.31 (q), 124.31, 126.22,
128.84, 130.96, 133.23, 133.76, 134.33, 136.10 (q), 143.37
(q), 153.93 (q), 174.49 (q); nmax (nujol) 3220, 2235, 1641,
1618, 1548, 1417, 1244, 1186 cm21;MS m/e (rel inten) 278
(M1, 34), 277 (100), 139 (7), 102 (8). Anal. calcd for
C15H10N4S: C, 64.72;H, 3.62;N, 20.12. Found: C, 65.02;
H, 3.61;N, 20.24.
4.6.3. Reaction of 8 with MnꢀOAc)3 in acetonitrile
ꢀTable 1, entry c). Following the general procedure in
Section 4.4.1, after 48 h column chromatography
afforded 2 (70%) together with traces of some unidenti®ed
products.

G. Calestani et al. / Tetrahedron 57 /2001) 7221±7233
7229
J1ˆJ2ˆ7.4 Hz, J3ˆ1.4 Hz), 7.65±7.80 (1.3H, bs1ddd,
J1ˆJ2ˆ7.3 Hz, J3ˆ1.5 Hz), 7.82±7.94 (2H, m), 8.05 (1H,
dd, J1ˆ7.9 Hz, J2ˆ0.9 Hz), 12.98 (0.7H, bs) (carboxylic
OH not apparent). This compound seems to show signi®cant
tautomerism of the NHCS moiety, since the signal at
12.98 ppm is clearly ascribable to less than one proton;
the remaining part appears as a very broad singlet at
,7.7 ppm, partially overlapped with the aromatic protons.
50 MHz 13C NMR (DMSO-d6) d 119.48, 119.92 (q),
128.05, 131.30, 132.15, 132.46, 134.57, 134.79, 136.60
(q), 139.28, 141.67 (q), 143.45 (q), 163.58 (q), 171.00 (q),
179.70 (q); nmax (nujol) 3396, 1680, 1652, 1538, 1406,
1202 cm21;MS m/e (rel inten) 297 (M1, 1), 279 (1), 263
(13), 253 (100), 235 (6), 145 (4), 134 (11), 92 (15), 90 (47),
64 (28). HRMS calcd for C15H11N3O2S 297.0572, found
297.0585], and 2-thioxo-2,3-dihydro-4/1H)-quinazolinone
(21, 4%), mpˆ295±3008C (identi®ed by comparison with
a commercial sample).
Figure 5. Molecular structure of 29, showing the atom labelling scheme.
4.6.4. Reaction of 8 with MnꢀOAc)3 and DEM in aceto-
nitrile ꢀTable 1, entry d). Following the general procedure
in Section 4.4.1, after 48 h column chromatography
afforded 2 (83%) together with traces of some unidenti®ed
products.
Compound 20 was identi®ed on the basis of the following
reasoning. The MS spectrum of 20 shows a very low
molecular ion (m/e 297 [,1]) but the two observed main
fragments (m/e 263 [13%] and 253 [100%]) de®nitely origi-
nate from it. Those peaks derive from the molecular ion by
loss of H2S and CO2, respectively. This was proved by a B/E
experiment, which also showed signi®cant loss of water
(m/e 279) from the M1. At the beginning of the MS analysis,
before vaporisation of the sample, thermal fragmentation
with loss of H2S occurs as a minor reaction since a clean
spectrum of a compound with M1 263 and main fragments
at m/e 235, 207, 145, and 90 (100%) appears at T,2008C;
further heating brings about volatilisation of the compound
with concomitant major loss of CO2. The base peak at m/e
253 in the MS spectrum and the decomposition observed
during melting point determination, together with the other
IR and NMR spectral data, clearly stand for the presence of
a carboxylic group. On the other hand, the facile loss of H2S
4.6.5. Reaction of 8 in acetonitrile/water in the absence of
MnꢀOAc)3 ꢀTable 1, entry e). Following the general pro-
cedure in Section 4.4.1, after 150 h column chromatography
separated compound 9 (40%) and 2 (trace amounts),
together with 33% of unreacted starting material.
4.6.6. Reaction of 8 in acetic acid/water in the absence of
MnꢀOAc)3. Following the general procedure in Section
4.4.1, after 48 h column chromatography yielded 2-[4-
oxo-2-thioxo-1,4-dihydro-3/2H)-quinazolinyl]benzonitrile
1
(19, 27%), mpˆ293±2968C [300 MHz H NMR (DMSO-
d6) d 7.20 (1H, bd, Jˆ8.0 Hz), 7.36 (1H, ddd, J1ˆ7.9 Hz,
J2ˆ7.5 Hz, J3ˆ0.9 Hz), 7.47 (1H, ddd, J1ˆ8.0 Hz, J2ˆ
1.1 Hz, J3ˆ0.5 Hz), 7.62 (1H, ddd, J1ˆJ2ˆ7.7 Hz, J3ˆ
1.1 Hz), 7.72 (1H, ddd, J1ˆ8.2 Hz, J2ˆ7.3 Hz, J3ˆ
1.3 Hz), 7.79 (1H, dd, J1ˆ7.8 Hz, J2ˆ1.6 Hz), 7.87 (1H,
bdd, J1ˆ7.4 Hz, J2ˆ1.6 Hz), 8.19 (1H, ddd, J1ˆ7.9 Hz,
J2ˆ1.3 Hz, J3ˆ0.3 Hz), 10.75 (1H, bs, NH);75 MHz 13C
NMR (DMSO-d6) d 112.59 (q), 115.95 (q), 116.47, 125.58,
128.05, 130.10, 131.09, 133.73, 135.05, 136.94, 140.04 (q),
142.03 (q), 160.08 (q), 175.63 (q) (one quaternary carbon
not apparent); nmax (CHCl3) 3400, 2220, 1710, 1625,
1610 cm21;MS m/e (rel inten) 279 (M1, 100), 278 (46),
246 (15), 220 (8), 162 (12), 145 (26), 119 (43), 90 (18).
Anal. calcd for C15H9N3OS: C, 64.50;H, 3.24;N, 15.04.
Found: C, 64.85;H, 3.23;N, 15.10]. The reaction also
afforded major amounts of an unidenti®able compound
with (apparent) m/e 278.
4.6.7. Hydrolysis of 8 in dioxane/HCl. A solution of 8
(1.11 g, 4 mmol) in a mixture of 1,4-dioxane (30 mL) and
10% aqueous hydrochloric acid (20 mL) was re¯uxed for
6 h. The mixture was then poured into an aqueous solution
of ammonium hydroxide and extracted with dichloro-
methane. The organic phase was separated and dried, the
solvent was evaporated and the residue chromatographed to
give 19 (20%), mpˆ293±2968C (spectral data identical to
those reported in Section 4.6.6), 2-[4-imino-2-thioxo-1,4-
dihydro-3/2H)-quinazolinyl]benzoic acid (20, 28%),
1
mpˆ278±2808C (dec.) [200 MHz H NMR (DMSO-d6) d
7.17 (1H, bs, imine NH), 7.38±7.50 (2H, m), 7.55 (1H, d,
Jˆ8.2 Hz), partially overlapped with 7.59 (1H, ddd,
Scheme 7. Fragmentation pattern of compound 20.

7230
G. Calestani et al. / Tetrahedron 57 /2001) 7221±7233
and H2O can only be explained by a structure in which the
carboxyl and thiocarbonyl groups are close enough to react
with each other in the way shown in Scheme 7.
(major), 135.13 (minor, H-2), 135.72 (q, minor), 139.11
(q, major), 139.78 (q, minor), 149.66 (q, major), 150.17
(q, minor), 154.48 (major, H-2), 157.74 (q, major);30 nmax
(CHCl3) 3455, 1720, 1670, 1620, 1600, 1570 cm21;MS m/e
(rel inten) 221 (M1, 38), 220 (100), 102 (8), 92 (8), 77 (7)],
As a further evidence, when 20 was allowed to react again
for several hours in a re¯uxing dioxane/conc. hydrochloric
acid solution, it afforded a mixture containing compound 34
as the major product. This was proved by MS and IR
analyses of the crude. The former revealed a compound
substantially identical in molecular ion and fragments to
that obtained from 20 at the beginning of the MS analysis
(M1 263);the latter showed an unambiguous new absorp-
tion at 1724 cm21 due to the new lactone carbonyl group of
34. These data are altogether clearly consistent with a struc-
ture still containing the aryl ring linked to the N-3 atom.
Therefore, they con®rm that no signi®cant rearrangement
occurred during the reaction under investigation.26
4-imino-3-phenyl-3,4-dihydro-2/1H)-quinazolinone
(11,
47%), mpˆ214±2188C (lit.11a mpˆ222±2238C) [300 MHz
1H NMR (DMSO-d6) d 6.50 (0.8H, bs), 7.08±7.17 (2H, m),
7.27±7.65 (5H, m), 8.12 (1H, d, Jˆ7.5 Hz), 8.85 (0.2H, bs),
11.03 (1H, bs); nmax (CHCl3) 3420, 3300, 1700, 1620 cm21
;
MS m/e (rel inten) 237 (M1, 18), 236 (100), 118 (10), 91
(8)], and 3-phenyl-2,4/1H,3H)-quinazolinedione (13, 8%),
mpˆ277±2798C (lit.11a mpˆ278±2818C).
4.7.3. Reaction of 30a. After 72 h column chromatography
separated 15 (5%), 2-/2-chloroanilino)-1,3-benzothiazole-
1
4-carbonitrile (3%), mpˆ175±1788C [300 MHz H NMR
(DMSO-d6) d 7.20±7.90 (6 H, m), 8.28 (1H, dd, J1ˆ8.0 Hz,
J2ˆ1.2 Hz), 11.00 (1H, bs, NH); nmax (CHCl3) 3410, 2210,
1630, 1590 cm21;MS m/e (rel inten) 287 (M112, 5), 285
(M1, 15), 250 (100), 235 (13), 207 (8), 90 (7).28 Anal. calcd
for C14H8ClN3S: C, 58.84;H, 2.82;N, 14.70. Found: C,
59.02;H, 2.81;N, 14.78], N-/2-chlorophenyl)-4-quinazo-
4.7. Reactions of the N,N0-diarylthioureas 10, 30a, and
30b with MnꢀOAc)3
4.7.1. General procedure. The general procedure described
for the reactions of isothiocyanates 1 and 3 with Mn(OAc)3
was used (Section 4.4.1). However, thioureas 10 and 30b
required three equiv. of Mn(III) instead of two for complete
disappearance of the starting material.
1
linamine (15%), mpˆ106±1098C31 [300 MHz H NMR d
7.08 (1H, ddd, J1ˆJ2ˆ7.7 Hz, J3ˆ1.4 Hz), 7.38 (1H, ddd,
J1ˆJ2ˆ7.9 Hz, J3ˆ1.4 Hz), 7.46 (1H, dd, J1ˆ8.2 Hz,
J2ˆ1.2 Hz), 7.61 (1H, ddd, J1ˆ8.2 Hz, J2ˆ7.0 Hz,
J3ˆ1.1 Hz), 7.82 (1H, ddd, J1ˆ8.3 Hz, J2ˆ7.1 Hz,
J3ˆ1.4 Hz), 7.92±7.98 (2H, m), 8.13 (1H, bs, NH), 8.80
(1H, dd, J1ˆ8.2 Hz, J2ˆ1.4 Hz), 8.84 (1H, s, H-2); nmax
(nujol) 3415, 1730, 1680, 1620, 1600 cm21;MS m/e (rel
inten) 257 (M112, 4), 255 (M1, 11), 254 (10), 220 (100),
102 (7), 92 (8). Anal. calcd for C14H10ClN3: C, 65.76;H,
3.94;N, 16.43. Found: C, 66.08;H, 3.92;N, 16.52. The
structure of this compound was con®rmed by X-ray diffrac-
tion (Fig. 6)], 3-/2-chlorophenyl)-4/3H)-quinazolinone
(6%), mpˆ160±1658C (lit.32 mpˆ1708C) [300 MHz 1H
NMR (DMSO-d6) d 7.65±7.77 (3H, m), 7.80±7.91 (3H,
m), 8.02 (1H, ddd, J1ˆ8.2 Hz, J2ˆ7.9 Hz, J3ˆ1.4 Hz),
8.31 (1H, dd, J1ˆ8.0 Hz, J2ˆ1.1 Hz), 8.43 (1H, s, H-2);
nmax (CHCl3) 1790, 1620, 1590 cm21;MS m/e (rel inten)
258 (M112, 2), 256 (M1, 7), 221 (100), 111 (9), 76 (6)],
and 3-/2-chlorophenyl)-4-imino-3,4-dihydro-2/1H)-quina-
zolinone (16%), mpˆ202±2048C [300 MHz 1H NMR
(DMSO-d6) d 6.61 (0.4H, bs, imine NH), 7.17±7.30 (2H,
m), 7.45±7.90 (5H, m1t, Jˆ8.0 Hz), 8.20 (1H, d,
Jˆ8.0 Hz), 9.00 (0.6H, bs, imine NH), 11.12 (1H, bs,
4.7.2. Reaction of 10. After 30 h column chromatography
separated 12H-[1,3]benzothiazolo[2,3-b]quinazolin-12-one
(15, 6%), mpˆ190±1928C (lit.27a mpˆ1938C) [the structure
of 15 was con®rmed by spectral comparison with an authen-
tic specimen prepared according to the literature:27a
1
300 MHz H NMR d 7.42±7.56 (4H, m), 7.64 (1H, dd,
J1ˆ7.5 Hz, J2ˆ1.5 Hz), 7.69 (1H, bd, Jˆ8.2 Hz), 7.81
(1H, ddd, J1ˆ8.4 Hz, J2ˆ7.3 Hz, J3ˆ1.5 Hz), 8.45 (1H,
dd, J1ˆ8.0 Hz, J2ˆ1.6 Hz), 9.05 (1H, dd, J1ˆ8.1 Hz,
J2ˆ1.6 Hz); nmax (nujol) 1785, 1595, 1580, 1555 cm21
;
MS m/e (rel inten) 252 (M1, 100), 224 (21), 134 (6), 90
(11)], 2-anilino-1,3-benzothiazole-4-carbonitrile (14, 8%),
mpˆ184±1868C [200 MHz 1H NMR d 7.24±7.40 (4H, m),
7.51±7.62 (3H, m), 7.66 (1H, dd, J1ˆ7.5 Hz, J2ˆ0.6 Hz),
8.10 (1H, dd, J1ˆ7.9 Hz, J2ˆ0.6 Hz);in CDCl the NH
3
proton gives a signal overlapped with the aromatic system,
whereas in DMSO-d6 it is signi®cantly deshielded
(,11 ppm);50 MHz 13C NMR d 105.62 (q), 117.27 (q),
123.87, 126.22, 129.06, 130.19, 130.66, 131.19, 134.63
(q), 139.86 (q), 148.50 (q), 171.83 (q); nmax (CHCl3)
3420, 2220, 1690, 1600, 1560 cm21;MS m/e (rel inten)
251 (M1, 100), 250 (44), 225 (9), 121 (8), 77 (16).28 Anal.
calcd for C14H9N3S: C, 66.91;H, 3.60;N, 16.72. Found: C,
67.21;H, 3.58;N, 16.82], N-phenyl-4-quinazolinamine (12,
amide NH); nmax (nujol) 3300, 3190, 1720, 1620 cm21
;
MS m/e (rel inten) 273 (M112, 2), 271 (M1, 4), 270 (7),
236 (100), 118 (4), 91 (11). Anal. calcd for C14H10ClN3O: C,
14%),
mpˆ215±2188C
(lit.11a
mpˆ220±2218C)29
[300 MHz 1H NMR (DMSO-d6) d 7.12 (0.9H, bt,
Jˆ7.0 Hz), 7.21 (0.1H, bt, Jˆ7.8 Hz), 7.30±7.50 (m),
7.62 (0.9H, ddd, J1ˆ7.8 Hz, J2ˆ7.2 Hz, J3ˆ1.2 Hz), 7.68
(0.1H, bt, Jˆ7.8 Hz), 7.75±7.95 (m), 8.52±8.62 (2H, s
[H-2]1d [Jˆ7.8 Hz] of the major isomer overlapped with
the corresponding signals of the minor one), 9.80 (0.9H, bs,
NH), 11.58 (0.1H, bs, NH);30 75 MHz 13C NMR (DMSO-d6)
d 114.29 (q, minor), 115.14 (minor), 115.16 (q, major),
122.44 (major), 122.96 (major), 123.70 (major), 126.17
(major), 127.53 (minor), 127.78 (major), 128.04 (minor),
128.41 (major), 128.74 (minor), 129.05 (minor), 132.94
Figure 6. Molecular structure of N-(2-chlorophenyl)-4-quinazolinamine,
showing the atom labelling scheme.

G. Calestani et al. / Tetrahedron 57 /2001) 7221±7233
7231
61.88;H, 3.70;N, 15.46. Found: C, 62.12;H, 3.67;N,
15.53].
ometer, u ±2u scan, scan width 1.210.34 tg u8, scan
speed 0.05±0.168 s21, graphite monochromated CuKa
Ê
radiation (lˆ1.54178 A). Pale-yellow prismatic crystal,
4.7.4. Reaction of 30b. After 48 h column chromatography
separated 15 (20%) and 4-imino-3-[2-/methylsulphanyl)-
phenyl]-3,4-dihydro-2/1H)-quinazolinethione (20%), mpˆ
224±2268C [MS m/e (rel inten) 299 (M1, 75), 253 (100),
266 (38), 252 (95), 237 (16), 102 (19);the melting point and
crystal size 0.48£0.30£0.28 mm, theta range for data
collection 4.60±69.988, index ranges 0#h#11, 0#k#19,
29#l#9, re¯ections collected 2309, independent re¯ec-
tions 2186 [R(int)ˆ0.0456], no absorption correction
performed, max. and min. transmission 1.0 and 0.931.
1
the H NMR spectrum of this compound are identical to
those obtained for the cyclisation product of 30b in
DMSO (see Section 4.5.4)]. The reaction also afforded
very minor amounts of unidenti®ed compounds.
Structure analysis and re®nement. The structure was solved
by direct methods with SIR9233 and re®ned by full matrix
least squares on F2 with SHELX93.34 The hydrogen atoms
were located in a DF map and re®ned isotropically. Data/
restraints/parameters 1837/0/209, goodness-of-®t on F2
0.853, ®nal R indices [I.2s(I)] R1ˆ0.0532, wR2ˆ0.1332.
Maximal D/sˆ0.000 for both non-hydrogen and hydrogen
atoms, maximum and minimum residual peak in the ®nal
difference Fourier map 0.379 and 20.249 e A23. The
molecular structure is shown in Fig. 5.
4.8. X-Ray crystal structure analysis of 6-thia-5,7,11c,12-
tetraazabenzo[e]aceanthrylene ꢀ2)
Crystal data. Empirical formula C15H8N4S, formula weight
Ê
276.32, temperature 293 K, wavelength CuKa (1.54178 A),
crystal system monoclinic, space group P21/c, unit cell
Ê
dimensions aˆ15.302(2), bˆ4.904(1), cˆ16.582(2) A,
Ê 3
bˆ107.46(5)8, volume 1187.0(5) A (by least squares
4.10. X-Ray crystal structure analysis of N-ꢀ2-chloro-
phenyl)-4-quinazolinamine
®tting of the setting angles of 28 automatically centred
re¯ections in the range 15.7#u#31.18), Zˆ4, density
(calculated)
1.5462 Mg/m3,
absorption
coef®cient
2.3112 mm21, F(000) 568.0.
Crystal data. Empirical formula C15H10Cl1N3, formula
weight 255.71, temperature 293 K, wavelength CuKa
Ê
(1.54178 A), crystal system monoclinic, space group P2/c,
Data collection and processing. Siemens AED diffract-
ometer, u ±2u scan, scan width 1.210.34 tg u8, scan
speed 0.05±0.168 s21, graphite monochromated CuKa
unit cell dimensions aˆ11.230(3), bˆ9.101(2),
Ê
Ê 3
cˆ13.173(3) A, bˆ112.91(2)8, volume 1240.1(6) A (by
least squares ®tting of the setting angles of 24 automatically
centred re¯ections in the range 20.4#u#38.68), Zˆ4,
density (calculated) 1.3696 Mg/m3, absorption coef®cient
2.6164 mm21, F(000) 528.0.
Ê
radiation (lˆ1.54178 A). Pale-yellow prismatic crystal,
crystal size 0.50£0.35£0.31 mm, theta range for data
collection 3.03±70.138, index ranges 218#h#17,
25#k#5, 27#l#20, re¯ections collected 2305, inde-
pendent re¯ections 2235 [R(int)ˆ0.0201], no absorption
correction performed, max. and min. transmission 1.0 and
0.886.
Data collection and processing. Enraf±Nonius CAD4
diffractometer, u ±2u scan, scan width 1.210.34 tg u8,
scan speed 0.05±0.168 s21, graphite monochromated CuKa
Ê
Structure analysis and re®nement. The structure was solved
by direct methods with SIR9233 and re®ned by full matrix
least squares on F2 with SHELX93.34 The hydrogen atoms
were located in a DF map and re®ned isotropically. Data/
restraints/parameters 2235/0/213, goodness-of-®t on F2
0.986, ®nal R indices [I.2s(I)] R1ˆ0.0501, wR2ˆ0.1602.
Maximal D/sˆ20.001 and 0.044 for non-hydrogen and
hydrogen atoms respectively, maximum and minimum
residual peak in the ®nal difference Fourier map 0.319
and 20.318 e A23. The molecular structure is shown in
Fig. 1.
radiation (lˆ1.54178 A). Pale-yellow prismatic crystal,
crystal size 0.45£0.22£0.18 mm, theta range for data
collection 4.27±69.988, index ranges 213#h#12,
0#k#11, 0#l#16, re¯ections collected 2448, independent
re¯ections 2350 [R(int)ˆ0.0191], no absorption correction
performed, max. and min. transmission 1.0 and 0.855.
Structure analysis and re®nement. The structure was solved
by direct methods with SIR9233 and re®ned by full matrix
least squares on F2 with SHELX93.34 The hydrogen atoms
were located in a DF map and re®ned isotropically. Data/
restraints/parameters 2217/0/204, goodness-of-®t on F2
0.853, ®nal R indices [I.2s(I)] R1ˆ0.0786, wR2ˆ0.2128.
Maximal D/sˆ0.000 and 20.001 for non-hydrogen and
hydrogen atoms respectively, maximum and minimum
residual peak in the ®nal difference Fourier map 0.523
and 20.573 e A23. The molecular structure is shown in
Fig. 6.
4.9. X-Ray crystal structure analysis of 11H-4-thia-
5,10,11c-triazacyclopenta[def]chrysen-11-one ꢀ29)
Crystal data. Empirical formula C15H7N3OS, formula
weight 277.30, temperature 293 K, wavelength CuKa
Ê
(1.54178 A), crystal system monoclinic, space group P21/
c, unit cell dimensions aˆ9.815(2), bˆ15.854(3),
Ê
Ê 3
Crystallographic data (excluding structure factors) for the
structures in this paper have been deposited with the
Cambridge Crystallographic Data Centre as supplementary
publication numbers CCDC 161826 (2), CCDC 161827
(29), and CCDC 161828 [N-(2-Chlorophenyl)-4-quinazo-
linamine]. Copies of the data can be obtained, free of
charge, on application to CCDC, 12 Union Road,
cˆ7.571(2) A, bˆ101.68(2)8, volume 1153.7(5) A (by
least squares ®tting of the setting angles of 24 automatically
centred re¯ections in the range 18.2#u#33.58), Zˆ4,
density (calculated) 1.5965 Mg/m3, absorption coef®cient
2.4217 mm21, F(000) 568.0.
Data collection and processing. Siemens AED diffract-

7232
G. Calestani et al. / Tetrahedron 57 /2001) 7221±7233
Cambridge CB2 1EZ, UK [fax: 144(0)-1223-336033 or
e-mail: deposit@ccdc.cam.ac.uk].
References
1. (a) Leardini, R.;Pedulli, G. F.;Tundo, A.;Zanardi, G.
J. Chem. Soc., Chem. Commun. 1984, 1320. (b) Leardini,
4.11. Semiempirical calculations
R.;Pedulli, G. F.;Tundo, A.;Zanardi, G. Synthesis 1985,
107. (c) Leardini, R.;Nanni, D.;Pedulli, G. F.;Tundo, A.;
Semiempirical calculations on radicals 28, 31, 32, and 33 as
well as the search for the reaction paths connecting 28±31
and 32±33 were carried out with the semiempirical program
included in the Wavefunction PC Spartan Pro 1.0.3 pack-
age. After a careful conformational search, the geometries
of the open-shell intermediates were fully optimised follow-
ing the MNDO-d parameterisation. A rough estimate of the
transition-state geometry was then located by calculating
the energy pro®le obtained by ®xing the geometries of the
starting and ®nal radical and varying in 20 steps the distance
between the two atoms involved in the reaction (ring open-
ing or ring closure). The transition state geometry and
energy were ®nally re®ned with the `Transition-State-
Geometry' option of the semiempirical routine.
Zanardi, G. J. Chem. Soc., Perkin Trans. 1 1986, 1591.
(d) Leardini, R.;Nanni, D.;Tundo, A.;Zanardi, G.
Gazz.
Chim. Ital. 1989, 119, 637. (e) Leardini, R.;Nanni, D.;Pedulli,
G. F.;Tundo, A.;Zanardi, G. J. Chem. Soc., Chem. Commun.
1989, 757. (f) Leardini, R.;Nanni, D.;Santori, M.;Zanardi, G.
Tetrahedron 1992, 48, 3961. (g) Nanni, D.;Pareschi, P.;
Rizzoli, C.;Sgarabotto, P.;Tundo, A. Tetrahedron 1995, 51,
9045. For a general review on the synthesis of heterocyclic
derivatives by radical cyclisation see: (h) Bowman, W. R.;
Bridge, C. F.;Brookes, P. J. Chem. Soc., Perkin Trans. 1
2000, 1.
2. (a) Saegusa, T.;Kobayashi, S.;Ito, Y. J. Org. Chem. 1970, 35,
2118. (b) Blum, P. M.;Roberts, B. P. J. Chem. Soc., Chem.
Commun. 1976, 535. (c) Blum, P. M.;Roberts, B. P. J. Chem.
Soc., Perkin Trans. 2 1978, 1313. (d) Bachi, M. D.;Balanov,
A.;Bar-Ner, N. J. Org. Chem. 1994, 59, 7752. (e) Bachi,
M. D.;Melman, A. J. Org. Chem. 1995, 60, 6242. (f) For a
general view on synthetic applications of radical addition to
isonitriles see: Nanni, D. Isonitriles: a Useful Trap in Radical
Chemistry. Radicals in Organic Synthesis;Renaud, P., Sibi,
M. P., Eds.;Wiley-VCH: Weinheim, 2001;Vol. 2, chapter
1.3, pp. 44±61.
Ring opening of 28 to 31 was found to be endothermic by
18 kcal/mol, with an activation barrier of 43.2 kcal/mol.
Structure 31 is the absolute minimum for the isothiocyanate
radical;the intermediate 32, which has a s-cis-like structure
and is therefore more suitable to study the subsequent ring
closure to 33, lies in a local minimum only 0.8 kcal/mol
higher than 31. For our purposes, the rotation barrier around
the C±NCS bond can be hence considered negligible.
Rotation of the C±amidine bond produces two virtually
degenerate minima, giving two structures with nearly
identical geometries in which the plane containing the
disubstituted aromatic ring contains the isothiocyanate
moiety as well and is orthogonal to the plane containing
the amidino group. The two amidine nitrogen atoms, both
of them bearing spin density, are therefore practically equi-
distant from the isothiocyanate carbon atom. Ring closure of
32 was calculated to be exothermic by 30.8 kcal/mol, with
an activation barrier of only 9.9 kcal/mol. The whole
rearrangement process (28!33) was hence found to be
exothermic by 12 kcal/mol.
3. (a) Barton, D. H. R.;Jaszberenyi, J. Cs.;Theodorakis, E. A.
Tetrahedron 1992, 48, 2613. (b) Lorenz, D. H.;Becker, E. I. J.
Org. Chem. 1963, 28, 1707. (c) Noltes, J. G.;Janssen, M. J. J.
Organomet. Chem. 1964, 1, 365. (d) John, D. I.;Tyrrel, N. D.
J. Chem. Soc., Chem. Commun. 1979, 345. (e) Barton, D. H. R.;
Bringmann, G.;Lamotte, G.;Motherwell, W. B.;Hay Mother-
well, R. S.;Porter, A. E. A. J. Chem. Soc., Perkin Trans. 1
1980, 2657. (f) John, D. I.;Tyrrell, N. D.;Thomas, E. J.
J. Chem. Soc., Chem. Commun. 1981, 901. (g) Witczak, Z. J.
Tetrahedron Lett. 1986, 27, 155. (h) Bachi, M. D.;
Denenmark, D. J. Org. Chem. 1990, 55, 3442. (i) Bachi,
M. D.;Melman, A. J. Org. Chem. 1995, 60, 6242. (j) Bachi,
M. D.;Melman, A. Synlett 1996, 60. (k) Bachi, M. D.;Bar-
Ner, N.;Melman, A. J. Org. Chem. 1996, 61, 7116. (l) For an
exhaustive discussion on the radical chemistry associated with
the thiocarbonyl group see: Crich, D.;Quintero, L. Chem. Rev.
1989, 89, 1413 (and references cited therein).
Energy values for radicals 28, 31, 32, and 33, and activation
barriers for ring opening of 28 and ring closure of 32 are
reported in Fig. 4, together with the optimised geometries.
All of the transition states are characterised by a single
imaginary vibrational frequency (376.8 and 364.2 cm21
for the transition states leading to 31 and 33, respectively)
resulting from a negative force constant in the diagonal form
of the Hessian;all of them collapse to the starting radicals
when optimised without the `Transition-State-Geometry'
option.
4. (a) Leardini, R.;Nanni, D.;Pareschi, P.;Tundo, A.;Zanardi,
G. J. Org. Chem. 1997, 62, 8394 (this paper also contains a
comprehensive literature citation on both chemical and spec-
troscopic properties of imidoyl radicals). (b) Camaggi, C. M.;
Leardini, R.;Nanni, D.;Zanardi, G. Tetrahedron 1998, 54,
5587. (c) Benati, L.;Leardini, R.;Minozzi, M.;Nanni, D.;
Spagnolo, P.;Zanardi, G. J. Org. Chem. 2000, 65, 8669.
5. One equiv. of Mn(OAc)3 acetate was added at the beginning
and another one after 18 h;after 38 h, the starting material was
completely consumed. Column chromatography separated a
®rst fraction containing elemental sulphur. When the reaction
was carried out at 808C we observed a considerable decrease
in the yield (ca. 20%): see Section 4. The starting isothio-
cyanate 1 was stable in the absence of manganese(III) acetate:
it was recovered almost quantitatively after prolonged warm-
ing at 608C both in acetic acid and acetonitrile, even in the
presence of a few equiv. of water.
Acknowledgements
The authors gratefully acknowledge ®nancial support
from CNR (Rome), MURST (2000-2001 Grant for
`Radical Processes in Chemistry and Biology: Syntheses,
Mechanisms, and Applications'), and the University of
Bologna (1999-2001 Funds for Selected Research
Topics).
6. Treatment of an enolisable compound with Mn(III) is a

G. Calestani et al. / Tetrahedron 57 /2001) 7221±7233
7233
È
22. Pinnow, J.;Sa mann, C. Chem. Ber. 1896, 29, 632.
23. Erdmann, H. Chem. Ber. 1899, 32, 3570.
well-known source of carbon radicals, see: Snider, B. B.
Chem. Rev. 1996, 96, 339.
7. The reaction also afforded small amounts (6%) of another
unidenti®ed compound with M1 209 (see Section 4). As
with isothiocyanate 1, compound 3 was perfectly stable in
the absence of Mn(OAc)3.
24. As a general rule, all the iminoquinazoline compounds (e.g.
30b) show, besides the characteristic peak of the thioamide
proton (12.5±13 ppm), two NMR absorptions at 6.5±7.2 and
8.8±9.4 ppm ascribable to the imine proton and probably due
to E/Z isomerism of the imine moiety (the sum of those two
signals always corresponds to one proton). On the contrary,
the IR spectra show only one signi®cant, generally sharp NH
band in the range 3220±3320 cm21. This behaviour was
observed with other analogous imines synthesised in our
laboratory (unpublished results). Compounds 8 and 20 seem
to exist as single geometric isomers.
8. Shipley Fry, H. J. Am. Chem. Soc. 1913, 35, 1539.
9. Korczynski, A. Bull. Soc. Chim. Fr. 1922, 31, 1184.
10. Ambiguous data are reported in the literature about the synthe-
sis and reactivity of N-(2-cyanophenyl)thioureas. However, it
seems that very mild heating is suf®cient to convert these
products into iminoquinazolines. See, for instance: (a) Taylor,
E. C.;Ravindranathan, R. V. J. Org. Chem. 1962, 27, 2622.
(b) Pazdera, P.;Ondracek, D.;Novacek, E. Chem. Papers
1989, 43, 771 In our reactions, the initial thiourea may form
by partial hydrolysis of 2 to 2-aminobenzonitrile, which is
continuously trapped by residual 2. This process is initiated
by the manganic salt, since 2 is stable in acetic acid in the
absence of Mn(OAc)3, but it is probably a non-radical step.
Interestingly, 2 is also stable in ethyl acetate in the presence of
potassium acetate, but it slowly gives major quantities of
2-aminobenzonitrile in the presence of catalytic amounts of
[18-crown-6]: this shows that, under suitable conditions,
acetate ions can bring about hydrolysis of the starting isothio-
cyanate.
25. Desulphurated products are usually always obtained in the
reactions of thioureas or their derivatives with Mn(OAc)3.
26. The crude also contained a minor product whose structure was
not fully established, but certainly included a cyano-group:
this was probably formed through hydrolytic fragmentation
of the N-3±C-4 bond. This is an additional proof in favour
of structure 20, since formation of a nitrile moiety should not
be expected from a rearranged ketimine derivative.
27. (a) Katz, L. J. Am. Chem. Soc. 1953, 75, 712. (b) Muthusamy,
S.;Ramakrishnan, V. T. Synth. Commun. 1992, 22, 519.
28. The formation of anilinobenzothiazoles was unquestionably
con®rmed by reacting N,N0-diphenylthiourea under the usual
conditions. The resulting 2-anilino-1,3-benzothiazole was
compared (mixed mp determination, spectral comparison,
and GC±MS analysis) with an authentic sample prepared
according to the literature (Hofmann, A. W. Chem. Ber.
1879, 12, 1126).
11. (a) Taylor, E. C.;Ravindranathan, R. V. J. Org. Chem. 1962,
27, 2622. (b) Perrin, D. D.;Pitman, I. H. Aust. J. Chem. 1965,
18, 763. (c) Perrin, D. D.;Pitman, I. H. J. Chem. Soc. 1965,
7071.
12. This reaction also afforded major quantities of unidenti®able
material: this could be the reason why the reactions performed
in acetic acid (Table 1, entries a and b) were less ef®cient than
those performed in acetonitrile (Table 1, entries c and d).
13. (a) Gomwalk, U. D.;McAuley, A. J. Chem. Soc. A 1968, 2948.
(b) Davies, G. Inorg. Chem. 1972, 11, 2488. (c) Turkevich,
N. M.;Dmitrishin, R. T. Visn. L'viv. Politekh. Inst. 1974, 82,
38 Chem. Abstr. 1975, 83, 52982e. (d) Dey, G. R.;Naik, D. B.;
Kishore, K.;Moorthy, P. N. J. Chem. Soc., Perkin Trans. 2
1994, 1625. (e) Kulkarni, M. J.;Rao, B. M. J. Indian Chem.
Soc. 1995, 72, 235. (f) Saha, S. K.;Greenslade, D. J. Indian J.
Chem. Technol. 1996, 3, 201.
29. Although compound 12 has been reported in several papers, its
NMR and MS spectral data are not available anywhere.
However, our spectral data, particularly the NMR absorptions
of the NH and H-2 protons, are consistent with those of
analogous compounds, see: Szczepankiewicz, W.;Suwinski,
J. Tetrahedron Lett. 1998, 39, 1785.
30. 1H NMR spectrum shows that this compound is a 1:9 mixture
of tautomers: the signal at 9.80 ppm is probably due to the
major, aromatic form (12, see Ref. 26), whereas that at 11.58
due to the ketimine form. The presence of two structures is
also evidenced by the 13C NMR spectrum (2 CH signals over-
lapped). The assignments were made on the basis of spectral
calculations, which perfectly predicted the consistent shift to
higher ®elds of the H-2 (more than 15 ppm) by passing from
structure 2 to the ketimine form.
14. (a) Leardini, R.;McNab, H.;Minozzi, M.;Nanni, D. J. Chem.
Soc., Perkin Trans. 1 2001, 1072. (b) Creed, T.;Leardini, R.;
McNab, H.;Nanni, D.;Nicolson, I. S.;Reed, D. J. Chem. Soc.,
Perkin Trans. 1 2001, 1079. See also Ref. 3l for the radical
chemistry of thiocarbonyl groups.
31. Gazit, A.;Chen, J.;App, H.;McMahon, G.;Hirth, P.;Chen, I.;
Levitzki, A. Bioorg. Med. Chem. 1996, 4, 1203 neither mp nor
spectral data reported.
15. Kice, J. L. Free Radicals;Kochi, J. K., Ed.;Wiley: New York,
1973;Vol. II, pp. 720±724.
32. Hurmer, R.;Vernin, J. Br. 1,093,977, 1967;( Chem. Abstr.
1968, 68, 39648w). This compound has also been reported
in: Gravier, D.;Dupin, J.-P.;Casadebaig, F.;Hou, G.;
Biosseau, M.;Bernard, H. Eur. J. Med. Chem. Chim. Ther.
1989, 24, 531. Daenens, P.;van Boven, M. Arzneim. Forsch.
1974, 24, 195;the former paper reports neither mp nor spectral
data, whereas the latter reports only a mass spectrum.
33. Altomare, A.;Cascarano, G.;Giacovazzo, C.;Guagliardi, A.;
Burla, M. C.;Polidori, G.;Camalli, M. J. Appl. Crystallogr.
1994, 27, 435.
16. Benati, L.;Camaggi, C. M.;Zanardi, G. J. Chem. Soc., Perkin
Trans. 1 1972, 2817.
17. Benati, L.;Montevecchi, P. C.;Tundo, A.;Zanardi, G. Gazz.
Chim. Ital. 1975, 105, 841.
18. (a) Gol'dfard, Ya. A.;Pokhil, G. P.;Belen'kii, L. I. Doklady
Akad. Nauk S.S.S.R. 1966, 167, 385. (b) Benati, L.;Camaggi,
C. M.;Zanardi, G. J. Org. Chem. 1975, 40, 966.
19. Pazdera, P.;Ondracek, D. Czech. CS 270,981, 1991; Chem.
Abstr. 1992, 117, 69590u, mp not reported.
20. Howard, J. C.;Klein, G. J. Org. Chem. 1962, 27, 3701.
21. Due to the very low solubility of compound 2 (even in DMSO-
d6), only the C±H signals are quoted.
34. Sheldrick, G. M. SHELX93, Program for Crystal Structure
È
Re®nement;University of Go ttingen, 1993.